Compressed air?

that might work indeed. but what kinds of volumes and pressures did they use?
to make it work well however we would need to make it work like a clockwork, as in that it won't release to fast early on and to slow later on, lucklily letting air out of a high pressure tank causes cold and eventually freezng, but this also means a liquid would then be send out and that would waste energy.
so we would need to either design and use a chamber which uses surrounding or tank heat to turn liquid air into gas like some kind of fuell explosion but constant and slower. or we would need to slow down the speed to keep the temps high enough or add in a heater.

also I think I just might have a idea involving some basic modern rocket science which despite fanboys claiming it to be hightech and very new, it in reality is around 100 year old or such, the inventor who used it just kind of treated it like something simple and logically anyone should expect to happen so didn't describe it to much seeing it more like basic physics(also turned out to be a invention people didn't use since they didn't understand how it worked(tesla turbine))
the tesla turbine (not to be confused with the current car company) actually is a insanely efficient and insanely simple and stable and small engine, kind of the perfect engine.
the engine has already been designed originally with aerospikes in it(now deemed to be recent innovations in rocketscience by many), as a funny extra the aerospike in that engine where actually more advanced than the ones in current day space engines.
one of the very interesting things about it is the self regulating/adapting/programmable aerospike parameters, you don't need to have a fixed design early on but you can change it in a analog way while it is running without any extra parts basically only depending on the rotation speed and centrifugal force.

what that means in simple words is that the faster such a turbine will spin the smaller it will make the nozzle formed by the aerospike, note that there is no physical nozzle, the nozzle is created out of air while it runs. (when I first realized this/figured this out and managed to prove it with a physical prototype I seemed to be the firt person after nikola tesla who actually realized this.).
but it means we can limit the flow of air making it more efficient, also if nothing is done it will kind of autobalance itself. it still will go faster with higher pressure but it prevents it from shooting the tank empty at once.

also some years ago I had designed a special part for using such a engine properly in a car or plane or such, as well as 2 other parts and one other system implementation based on existing parts.
these engines will be much cheaper and simpler to make and maintain as they can be practically maintainance free other than the bearings perhaps but there is only 1 moving part and that is just a axis with some disks on it meaning no vibration or such, which also allows to make the vehicles a lot more leightweight.
recently I also just happened to find a new part for it, even though if we are going to run it on pressurized air that won't be needed anymore.